中文摘要：

本文的主要工作是对季节调整中结构分量模型的选择及稳健性问题进行研究.通过构造备选模型、引入正态先验分布及对MCMC抽样方案进行重新设计,本文提出了能同时有效地辨别出分量个数和分量随机性与否的模型选择方法.将该方法应用于对中国季度GDP序列的季节调整建模分析中.分析结果显示,本文所提出的模型选择方法具有较高的筛选能力,并且对先验分布超参数值的变动也具有很好的稳健性.

英文摘要：

The main work of this paper is to study the selection and robustness of Structural Component model in Seasonal Adjustment. Through constructing the alternative model, employing a normal prior distribution, and deriving a new MCMC sampling scheme, a new model selection method that can distinguish which components to include in the model and whether these components are fixed or time-varying is developed in this paper. Finally, we apply this method to analyze the seasonal series of China＇ s GDP. The results show that the method we employ has a superior performance, and robust to the prior＇ s hyperparameter.